Interaction of Dietary Patterns and Physical Activity with Low Back Pain in Pre- to Post-Menopause: A Cross-Sectional Study Authors: Jing Zhao, Haiwen Lu, Sheng Wang, Rixin Niu, Xinyan Wang, Pengpu Wang, Hailin Li, Yanling Wang, Huiqiu Zheng, Yanling Li, Xin Liu, Yingjie Tian, Qianqian Du, Bowen Zhou, Jingxia Fan, Xuemei Wang Published: 2026 ABSTRACT AbstractBackgroundLow back pain (LBP) in perimenopause and postmenopausal women is an important public health concern because of its high prevalence and associated social and economic burden. Dietary patterns and physical activity have attracted increasing attention as modifiable determinants of LBP, but the impact of their interaction, particularly across menopausal stages, has not been well characterized.MethodsThis cross-sectional analysis included 1530 adult women from the Ordos Resident Health White Paper Program. Dietary patterns were derived using principal components analysis of food frequency questionnaire data. Physical activity was assessed using the International Physical Activity Questionnair. LBP was defined as self-reported pain, ache or discomfort in the lumbosacral region (Verbal Rating Scale score 1–3). Logistic regression models were used to examine associations between dietary patterns, physical activity and LBP. Interaction between dietary patterns and physical activity was assessed overall and stratified by menopausal status.ResultsThe prevalence of LBP was 12.6%. Three major dietary patterns were identified: a plant-based dietary pattern (PBDP), a northern pastoral dietary pattern (NPDP) and a northern traditional dietary pattern (NTDP). Only PBDP was significantly associated with LBP. Compared with women in the highest quartile of adherence to PBDP (Q4), those in the lowest quartile (Q1) had higher odds of LBP (adjusted OR = 2.879, 95% CI: 1.749–4.739). No significant association with LBP was observed for NPDP. In joint analyses, the highest odds of LBP were observed among women with higher-intensity physical activity and non-adherence to PBDP (Q1) (adjusted OR = 3.777, 95% CI: 1.991–7.166). In analyses stratified by menopausal status, the risk associated with higher-intensity physical activity and non-adherence to PBDP increased from perimenopausal to postmenopausal women, with adjusted ORs of 3.600 (95% CI: 1.029–12.592) and 5.752 (95% CI: 2.189–15.114), respectively.ConclusionsIn this population of midlife women, higher-intensity physical activity was associated with increased risk of LBP, and this risk rose from perimenopause to postmenopause. Adherence to PBDP attenuated the LBP risk associated with higher-intensity physical activity, with the beneficial effect becoming more pronounced after menopause. These findings suggest that menopause-specific combinations of dietary improvement and appropriate physical activity may help mitigate the burden of LBP in perimenopausal and postmenopausal women.Keywords: Low back pain, Perimenopause and postmenopausal women, Dietary pattern, Plant-based diet, Physical activity, Interaction Introduction Low back pain (LBP) is one of the most common musculoskeletal complaints and a leading cause of disability and reduced quality of life worldwide [1]. Its prevalence increases with age and is particularly high among women in midlife and older age [2]. In addition to individual suffering, LBP imposes a substantial economic burden through health care use and loss of productivity. Although most cases of LBP are non-specific and do not have a single identifiable pathological cause, a range of potentially modifiable lifestyle factors, including physical activity and diet, have been implicated in its occurrence and persistence [3, 4]. The menopausal transition is accompanied by hormonal changes, alterations in body composition and functional decline that may influence the risk of LBP. Declining oestrogen levels are associated with reductions in bone mass and muscle strength, as well as with changes in fat distribution and low-grade systemic inflammation. These changes may increase mechanical stress on the spine and surrounding structures and may lower the threshold for pain. However, not all perimenopausal and postmenopausal women develop LBP, suggesting that lifestyle factors, such as physical activity and dietary patterns, may modify individual susceptibility.Both physical activity and diet are key components of a healthy lifestyle and are recommended for the prevention and management of many chronic conditions. Regular physical activity can improve muscle strength [5], balance and functional capacity [6, 7], but excessive or improperly performed higher-intensity activity may exacerbate musculoskeletal pain in vulnerable individuals [8]. In postmenopausal women, muscle strength can often be maintained only by limited types of low-intensity physical activities [9], whereas higher-intensity physical activity may contribute to a rapid decline in lower limb muscle strength [10]. A UK Biobank study showed that postmenopausal women had a weaker bone response to the same level of higher-intensity habitual physical activity than premenopausal women [11]. Similarly, dietary patterns characterized by higher intakes of plant-based foods, high-quality protein and anti-inflammatory nutrients have been associated with better musculoskeletal health and reduced chronic pain in some studies. Higher protein intake has been shown to help preserve muscle mass and may improve LBP [12]. In real-world settings, people consume complex combinations of foods rather than isolated nutrients, and dietary components may interact to produce synergistic effects [13]. Plant-based protein sources may improve LBP by preserving muscle mass and strength [14]. For example, higher adherence to the Mediterranean diet, characterized by a higher daily intake of whole grains, fruits, vegetables, tree nuts, legumes, and olive oil and a moderate intake of fish and poultry–has been associated with improved muscle strength [15]. Among perimenopausal women, adherence to a plant-rich dietary pattern with vegetables, legumes and fruits has been reported to benefit musculoskeletal health [16]. In postmenopausal women, healthy (fruit and vegetable) or prudent (nuts, mushrooms, algae, and seafood) dietary patterns has been associated with a reduced risk of musculoskeletal disorders [17]. Collectively, these findings suggest that diet and physical activity can modify muscle mass and strength and thus may improve or worsen LBP in perimenopausal and postmenopausal women. Physical activity and diet may also act synergistically on lumbar spine bone mineral density (BMD) [18]. From this perspective, the musculoskeletal damage potentially induced by higher-intensity physical activity and the protective effects of favourable dietary patterns represent key lifestyle factors that could modulate LBP risk, especially in midlife women. However, most previous studies have considered diet and physical activity separately, and evidence on how overall dietary patterns and physical activity jointly relate to LBP in perimenopausal and postmenopausal women remains limited, particularly in populations from northern China [19, 20]. Therefore, in this study, we aimed to evaluate the combined effects of dietary patterns and physical activity on LBP in women, with a particular focus on differences between perimenopause and post menopause. By characterizing these interactions, our findings may help inform more rational combinations of dietary patterns and physical activity to address LBP in perimenopausal and postmenopausal women. Materials and methods Data source and subjects Participants were recruited from adults (aged ≥ 18 years) enrolled in the Ordos Resident Health White Paper Program. For the present analysis, we included women with complete dietary, physical activity and LBP data. We excluded individuals with conditions that might substantially alter their diet, including coronary heart disease, kidney disease, cancer, and brucellosis. Individuals with daily energy intake out of the range of 800–4200 kcal/day were also excluded, as were those with missing data on key variables. After applying these criteria, 1530 female participants were included. The Biomedical Research Ethics Committee of Inner Mongolia Medical University approved the study protocol (No. YKD202301132). All participants provided written informed consent before the start of the investigation. Dietary data were collected using food frequency questionnaire (FFQ) data. Participants were asked about their dietary intake during the preceding 30 days. The FFQ covered fruits; vegetables, mushrooms, and algae; beans and their products; eggs and egg products; coarse cereals; aquatic products; beef and mutton; tea and coffee drinks, and other beverages; poultry, other meat products, and animal organs; snacks, vegetarian desserts, nuts and fried dough foods; alcoholic drinks; milk and dairy products; tubers; wheat flour and its products; rice and its products; salt and pickled vegetables; pork; cooking oil, condiment. For each item, participants reported whether it was consumed, the frequency of consumption (times/day, times/week, times/month, times/year), and the average amount per occasion (g). Standardized dietary survey mapping and food models were used to assist recall. Outcome definitions Pain was categorized according to the Verbal Rating Scale (VRS) [21]. 0 = no pain, 1 = mild pain, 2 = moderate pain, and 3 = severe pain. Because our focus was on the presence of pain rather than differences in intensity, categories 1–3 were combined as pain and category 0 was defined as no pain (Table 1). In this study, LBP specifically refers to self-reported pain, ache or discomfort in the lumbosacral region, as indicated on a body diagram, that had occurred at least once during the previous 12 months. We did not assess pain in other anatomical regions or multi-site pain, and we did not collect information on pain duration or LBP related disability. Table 1Pain classificationScorePainFeatures0No painNo discomfort or pain1Slight painThe patient could be distracted from feeling pain, and no analgesia was required2Moderate painThe patient felt moderate pain even while concentrating on some other activity, and an analgesic was required3Severe painThe patient could no longer perform any activity and needed to lie down and rest (analgesics had little or no effect on pain relief) Pain classification Other definitions Other variables were assessed using questionnaires and physical examinations, including age [22] (premenopausal women 18–44 years, perimenopausal women 45–55 years, and postmenopausal women ≥ 56 years); nationality (Han, Mongolian, and other ethnic groups); area (urban, rural, and pastoral); socioeconomic status (SES; low, medium, and high), sleep deficit (yes, no), body mass index (BMI), BMD (T-scores), sedentary time (≤ 2.00 h, 2.01–3.00 h, 3.01–5.00 h, > 5.01 h), and physical activity. BMI was calculated as weight(kg) divided by height(m2). Physical activity Physical activity was assessed using the International Physical Activity Questionnaire (IPAQ) and converted to weekly MET-minutes of moderate to vigorous physical activity [23]. Total physical activity (MET-min/week) was calculated according to the standardized IPAQ scoring protocol. Initially, participants were classified into low (0–1908 MET-min/week), moderate (1909–5129 MET-min/week) and high (≥ 5130 MET-min/week) physical activity categories. In preliminary analyses, the associations of moderate and high physical activity with LBP were similar in direction and magnitude. For parsimony and to increase statistical power in stratified and interaction analyses, we therefore combined the moderate and high categories and defined them as “higher-intensity physical activity”. Statistical analysis Continuous variables were expressed as mean ± standard deviation SD and compared between groups using the t-test. Categorical variables were expressed as numbers and percentages and were analyzed using the χ2 test. Principal component analysis(PCA) was performed to identify the major dietary patterns by generating standardized factor scores for each pattern. Kaiser-Meyer-Olkin (KMO) values and Bartlett’s test of sphericity were used to assess sampling adequacy and factorability. Orthogonal(varimax) rotation was then performed to maximize variance and minimize cross-loading between factors, improving the explanatory power of the factor model. Model selection was guided by scree plots, eigenvalues (> 1), and interpretability of each factor. Food groups with factor loadings ≥ 0.2 were considered major contributors to each pattern. The cumulative variance explained by these factors was used to capture variability in food intake. Each dietary pattern was labelled according to the food group with the highest factor loading and used primarily for descriptive interpretation. Because dietary pattern factor scores are continuous variables that may have non-linear relationships with outcomes, categorization into quartiles (Q1–Q4) allowed us to evaluate potential non-linear trends and to compare extreme categories in a way that is easier to interpret clinically. Higher scores (Q4) indicated adherence to the pattern, whereas lower scores (Q1) indicated non-adherence. Multivariable logistic regression analyses were performed to determine the relationships between LBP and the three dietary patterns, using unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (CIs). Model 1 was adjusted for age, nationality, area, SES, BMI, BMD (T-scores), sleep deficit, and sedentary time. Model 2 further adjusted for physical activity. The highest quartile (Q4) was used as the reference category. Binary logistic regression models were then used to investigate how the interaction between dietary patterns and physical activity level influenced the risk of LBP. The median of each dietary pattern quartile was treated as a continuous variable to explore linear trends. Statistical analyses were performed using IBM SPSS 25.0 (IBM Corp., Armonk, NY, USA). R 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria) was used for plotting. A two-sided P < 0.05 was considered significant. Results Participant characteristics Among the 1530 women included in this study, 12.6% (193/1530) had LBP. Overall, 87.1% were Han nationality, 40.1% were postmenopausal women, 72.1% resided in rural or pastoral areas, 21.0% had high SES, 28.0% reported sleep deficit, and 50.4% engaged in higher-intensity physical activity. The mean BMI was 25.67 (SD, 3.87) kg/m2 and the mean BMD (T-scores) was − 1.99 (SD, 1.83). Significant differences were observed between women with and without LBP in age, area, SES, BMI, BMD (T-scores), physical activity, and sedentary time (all P < 0.05), whereas nationality and sleep deficit did not differ significantly. (Table 2). Table 2Participant characteristics in low back pain and non-low back pain groupsVariablesTotal, n(%)low back pain, n(%)t/χ2 P Age 14.544< 0.05 Premenopausal women (18–44 years)364(23.8)32(8.8) Perimenopausal women (45–55 years)552(36.1)60(10.9) Postmenopausal women (≥ 56 years)614(40.1)101(16.4) Nationality 0.4810.488 Han1332(87.1)165(85.5) Mongolian and other ethnic groups198(12.9)28(14.5) Area 4.1480.042 Urban427(27.9)42(21.8) Rural or pastoral1103(72.1)151(78.2) SES 16.429< 0.05 Low507(33.1)79(40.9) Middle701(45.8)94(48.7) High322(21.0)20(10.4) BMI, Mean(SD) 25.67 (3.87)26.23 (3.73)−2.1730.030 BMD (T-scores), Mean(SD) −1.99 (1.83)−2.26 (1.79)2.2810.023 Sleep deficit 2.9490.086 Yes428(28.0)64(33.2) No1102(72.0)129(66.8) Physical activity 3.8520.050 Low-intensity759(49.6)83(43.0) Higher-intensity771(50.4)110(57.0) Sedentary time 17.7090.001 <=2.00754(49.3)120(62.2) 2.01–3.00.01.00270(17.6)33(17.1) 3.01–5.00.01.00269(17.6)22(11.4) 5.01+237(15.5)18(9.3) Participant characteristics in low back pain and non-low back pain groups Dietary patterns were defined based on the food groups with factor loadings ≥ 0.2. The three dietary patterns were identified: PBDP (fresh vegetables, fruits, beans and bean products, eggs and egg products, aquatic products), NPDP (beef, mutton, milk and dairy products), NTDP (rice and its products, salt and pickled vegetables, pork, cooking oil). (Tables A1 and A2, Figure A1). Significant differences (P < 0.05) in the intake of fruits, vegetables and mushrooms/algae, beans and their products; eggs and egg products; coarse cereals; and aquatic products were observed between women with and without LBP, with lower intakes in the LBP group (Table 3). Table 3Comparison of daily dietary intakes between women with and without LBP (g/d or ml/d)Food groupLBP(N = 193)no LBP(N = 1337)t P Fruits87.39 ± 91.83139.76 ± 125.745.575< 0.05Vegetable mushrooms and algae187.57 ± 160.01223.64 ± 149.423.106< 0.05Beans and their products43.04 ± 53.7058.91 ± 78.942.703< 0.05Eggs and egg products35.04 ± 29.9041.59 ± 33.422.579< 0.05Coarse cereals59.23 ± 72.9377.81 ± 98.872.515< 0.05Aquatic product10.52 ± 21.9217.58 ± 37.172.574< 0.05Beef and mutton57.91 ± 77.9663.53 ± 85.590.8620.389Other beverages18.31 ± 69.2019.04 ± 71.710.1340.894Tea and coffee drinks272.40 ± 498.46242.48 ± 525.37−0.7440.457Poultry, other meat products, and animal organs37.50 ± 62.1851.75 ± 69.142.71< 0.05Snacks, vegetarian desserts, nuts145.44 ± 112.75168.66 ± 138.722.2210.183Fried dough foods7.60 ± 21.118.22 ± 19.220.4150.678Alcoholic drinks0.84 ± 5.380.95 ± 9.910.160.873Milk and dairy products174.14 ± 187.39174.50 ± 196.460.0240.981Tubers53.91 ± 60.3862.91 ± 56.532.049< 0.05Wheat flour and its products49.75 ± 49.0056.12 ± 57.061.4750.141Rice and its products90.70 ± 63.2088.82 ± 61.71−0.3940.693Salt and pickled vegetables24.75 ± 37.0122.75 ± 30.18−0.8350.404Pork72.66 ± 58.7187.31 ± 88.602.227< 0.05Cooking oil39.24 ± 25.4442.89 ± 38.281.2840.199Condiment13.02 ± 11.1714.24 ± 22.010.750.453 Comparison of daily dietary intakes between women with and without LBP (g/d or ml/d) In the group with PBDP, significant differences were observed in the nationality, area, SES, BMI, BMD (T-scores), physical activity, and sedentary time among the subgroups of different dietary patterns (P < 0.05). The LBP prevalence was lower in women adhering more to PBDP (P < 0.05). In NPDP group, significant differences were observed in age, nationality, area, SES, sleep deficit, physical activity, and sedentary time among the subgroups of different dietary patterns (P < 0.05). In NTDP group, significant differences were observed in BMI, nationality, area, SES, sleep deficit, and physical activity among the subgroups of different dietary patterns (P < 0.05). (Tables A3-1, A3-2, and A3-3 in Appendix A). Association between dietary pattern and LBP Table 4 shows the associations between dietary patterns and LBP. For PBDP, compared with women in Q4, those in Q1 had the highest risk of LBP. In the crude model, the OR for Q1 versus Q4 was 3.426 (95% CI: 2.161–5.432). After adjustment for age, nationality, area, SES, BMI, BMD (T-scores), sleep deficit and sedentary time (Model 1), the OR decreased to 2.921 (95% CI: 1.776–4.804) but remained significant. After further adjustment for physical activity (Model 2), the OR for Q1 versus Q4 was 2.879 (95% CI: 1.749–4.739). A significant inverse trend across quartiles was observed (P for trend < 0.05). Table 4Multivariable-adjusted odds ratios and 95% confidence intervals for LBP across categories of three dietary patternsMajor dietary patternCategoriesCrudeModel 1aModel 2bPBDPQ4111Q31.921(1.179, 3.130)*1.795(1.087, 2.965)*1.772(1.073, 3.927)*Q22.990(1.898, 4.712)*2.738(1.707, 4.393)*2.706(1.686, 4.343)*Q13.426(2.161, 5.432)*2.921(1.776, 4.804)*2.879(1.749, 4.739)*P- trend< 0.05< 0.05< 0.05NPDPQ4111Q30.773(0.449, 1.328)0.843(0.471, 1.510)0.865(0.483, 1.550)Q21.069(0.655, 1.744)1.073(0.614, 1.875)1.112(0.635, 1.948)Q11.182(0.746, 1.872)1.099(0.639, 1.890)1.150(0.667, 1.984)P- trend0.2990.7130.663NTDPQ4111Q31.319(0.785, 2.216)1.465(0.865, 2.482)0.597(0.360, 0.992)*Q21.834(1.136, 2.963)*2.046(1.255, 3.336)*0.883(0.569, 1.372)Q11.393(0.855, 2.270)1.597(0.966, 2.641)1.241(0.837, 1.838)P- trend0.0780.0350.030aModel 1 adjusted for age, nationality, area, SES, BMI, BMD (T-scores), sleep deficit, sedentary timebModel 2 adjusted for variables in model 1, physical activity*P < 0.05 Multivariable-adjusted odds ratios and 95% confidence intervals for LBP across categories of three dietary patterns aModel 1 adjusted for age, nationality, area, SES, BMI, BMD (T-scores), sleep deficit, sedentary time bModel 2 adjusted for variables in model 1, physical activity *P < 0.05 For NPDP, no statistically significant association with LBP was observed in crude or adjusted models; ORs were close to 1 and P for trend values were not significant. For NTDP, crude analyses suggested higher LBP risk in some lower quartiles versus Q4, but this pattern weakened after multivariable adjustment. In Model 2, only Q3 showed a significantly lower odds compared with Q4 (OR = 0.597, 95% CI: 0.360–0.992), whereas other quartiles were not significantly different. Overall, PBDP was the only dietary pattern consistently associated with LBP. Interaction between physical activity and PBDP on LBP The joint association of PBDP and physical activity with LBP is presented in Fig. 1. Using women with low-intensity physical activity and the highest adherence to PBDP (Q4) as the reference, we found that non-adherence to PBDP (Q1) was associated with substantially higher odds of LBP in both physical activity categories. Models are adjusted for age, nationality, area, socioeconomic status, BMI, BMD (T-scores), sleep deficit and sedentary time. Among women with low-intensity physical activity, the adjusted ORs (95% CIs) for LBP were 2.207 (1.093–4.458) for Q1, 2.544 (1.332–4.860) for Q2 and 1.677 (0.830–3.390) for Q3, compared with Q4. Among women with higher-intensity physical activity, the corresponding adjusted ORs were 3.777 (1.991–7.166) for Q1, 2.886 (1.537–5.421) for Q2, 1.984 (1.019–3.862) for Q3 and 0.962 (0.460–2.012) for Q4. These results indicate that higher-intensity physical activity is associated with an elevated risk of LBP, particularly among women who do not adhere to PBDP. Greater adherence to PBDP appears to attenuate the LBP risk associated with higher-intensity physical activity. Multivariable-adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for LBP according to quartiles of adherence to the PBDP, stratified by physical activity level. The reference group is women with low-intensity physical activity in the highest quartile of PBDP (Q4, OR = 1.00). Models are adjusted for age, nationality, area, SES, BMI, BMD (T-scores), sleep deficit and sedentary time Subgroup analysis by menopausal status Subgroup analyses were conducted according to menopausal status (Fig. 2). Among perimenopausal women with higher-intensity physical activity, non-adherence to PBDP (Q1) was associated with an adjusted OR of 3.600 (95% CI: 1.029–12.592) for LBP. Among postmenopausal women with higher-intensity physical activity, the corresponding OR was 5.752 (95% CI: 2.189–15.114). Among women with low-intensity physical activity, non-adherence to PBDP (Q1) was also associated with increased LBP risk (OR = 2.893, 95% CI: 1.042–8.032). Overall, these findings suggest that the combined effect of higher-intensity physical activity and non-adherence to PBDP (Q1) on LBP risk becomes more pronounced from perimenopause to postmenopause, and that adherence to PBDP may mitigate this risk across menopausal stages. Subgroup logistic regression estimates of the joint association of PBDP adherence and physical activity with LBP by menopausal status. The reference group is women with low-intensity physical activity in the highest quartile of PBDP (Q4, OR = 1.00). Models are adjusted for age, nationality, area, SES, BMI, BMD (T-scores), sleep deficit and sedentary time Discussion In this cross-sectional study of women from Ordos, China, we observed three main findings. First, among the three identified dietary patterns, only the PBDP was consistently associated with LBP, with lower adherence linked to higher odds of LBP. Second, higher-intensity physical activity was associated with an increased risk of LBP, and this association appeared stronger in postmenopausal than in perimenopausal women. Third, diet and physical activity interacted: women with higher-intensity physical activity and low adherence to PBDP had the highest risk of LBP, whereas adherence to PBDP appeared to attenuate this risk. We also found that the overall prevalence of LBP in this population was 12.6%, which is higher than the global prevalence of 8.7% in adult women [24], but lower than a reported prevalence of 29.6% among Chinese women aged ≥ 45 [25] years. Previous studies have reported LBP prevalences of 22.5% in perimenopausal women and 13.0% in menopausal women in China [25, 26], indicating substantial heterogeneity across populations. Consistent with those reports, our data show that LBP is common in midlife and older women. Age-appropriate physical activity is crucial [27]. In general, physical activity improves muscle health and supports bone density [28–30]. However, excessive or poorly adapted higher-intensity physical activity can induce intense local and nerve root pain in the intervertebral disc [31] and may be harmful for individuals with reduced bone and muscle reserves. In our study, the LBP risk associated with higher-intensity physical activity increased from perimenopausal to postmenopausal women. This is biologically plausible. Bone loss in perimenopausal and postmenopausal women is largely related to oestrogen deficiency due to ageing and ovarian failure [32]. Hormone-induced bone loss is thought to increase fragility fracture risk and may contribute to pain [33]. In perimenopausal women, rapid declines in oestrogen accelerate spinal degeneration and increase the incidence of lumbar spondylolisthesis [34]. Hormone deficiency and/or higher-intensity physical activity may also cause microdamage and delay bone recovery [35]. Skeletal muscle strength declines with age [36]. In postmenopausal women, hormonal changes further accelerate bone loss [37] and the decline of muscle mass and strength [38]. During the perimenopausal transition, hormonal changes may outweigh the beneficial effects of physical activity on bone mass and muscle strength, as bone tissue is strongly regulated by sex hormones [29]. Under these conditions, certain types of higher-intensity physical activity may be insufficient to reverse muscle atrophy and may instead exacerbate musculoskeletal damage [30, 39]. Physical inactivity also tends to increase from perimenopause to postmenopause [40], which may make occasional bouts of higher-intensity activity particularly injurious. These mechanisms together may explain the stronger adverse impact of higher-intensity physical activity on LBP in postmenopausal women observed in our study. Our finding that PBDP was associated with lower LBP risk is consistent with evidence linking plant-rich dietary patterns to better musculoskeletal outcomes. PBDP in this study was characterised by higher intakes of vegetables, fruits, beans, eggs and aquatic products. This pattern shares similarities with the Mediterranean diet, which is rich in fruits, vegetables and fish and has been shown to enhance muscle strength and function in older adults [41]. A plant-based diet can adequately meet protein requirements in older adults, and its abundant antioxidants and anti-inflammatory nutrients may reduce the risk of muscle loss [42]. Although some studies emphasise the rapid anabolic effects of animal protein, plant-based protein sources may have unique advantages for long-term musculoskeletal health and inflammatory balance [43]. Given that skeletal muscle mass decreases with ageing, diet—especially protein intake—is one of the most important modifiable factors influencing muscle mass in middle-aged and older adults [44]. We also observed that the beneficial effects of dietary improvement on LBP appeared to become more pronounced from perimenopause to postmenopause. This may be related to the accelerated loss of muscle mass [45] and increased systemic inflammation [46] associated with declining estrogen levels, which may heighten the body’s sensitivity to nutritional interventions [47]. For perimenopausal women, LBP prevention may require both adapting physical activity and improving skeletal muscle mass through increased protein intake to compensate for exercise-induced muscle protein loss [48]. For postmenopausal women, higher plant-based protein intake is a protective factor against muscle loss due to ageing and physical activity [49], and increasing plant protein intake may mitigate bone loss and reduce musculoskeletal risk [50]. Taken together, these findings support the hypothesis that the dietary components of PBDP help maintain skeletal muscle mass and thereby improve LBP. This study has several strengths. We used detailed FFQ data to derive dietary patterns, allowing us to capture real-world eating behaviors rather than isolated nutrients. We also considered physical activity and sedentary time and performed stratified analyses by menopausal status, providing a more nuanced understanding of how lifestyle factors interact to influence LBP in midlife women. However, several limitations should be acknowledged. First, as a cross-sectional study, we cannot establish causal relationships between PBDP adherence, physical activity and LBP. We also could not determine the direction of causality or identify the optimal intensity and type of physical activity for LBP prevention. Reverse causation is possible; women with LBP may modify their diet or physical activity because of pain. Prospective cohort studies and randomized controlled trials are needed to clarify temporal relationships and causality. Second, dietary intake was assessed using an FFQ, which is prone to recall bias and measurement error. Such misclassification is likely non-differential with respect to LBP and may have biased associations towards the null. Future research could use repeated 24-hour dietary recalls or food diaries to improve accuracy. Third, although we adjusted for a broad range of sociodemographic and lifestyle factors, residual confounding from unmeasured variables (e.g. detailed occupational ergonomics, psychosocial stressors, pain coping strategies) cannot be excluded. Fourth, we assessed only the presence versus absence of LBP in the lumbosacral region using the VRS and did not collect detailed information on pain duration, recurrence, specific underlying diagnoses or LBP-related disability. Therefore, our outcome likely reflects predominantly non-specific LBP, and we could not differentiate between acute and chronic or disabling LBP. The mechanistic pathways discussed with regard to menopause, bone, muscle and pain are therefore hypothetical, based on previous literature, and should not be interpreted as causal mechanisms directly demonstrated by our data. Our findings underscore the importance of integrating both dietary quality and appropriately prescribed physical activity into community-based strategies for the prevention and management of LBP in women. In particular, the increased risk of LBP among postmenopausal women with low adherence to a PBDP and higher physical activity levels suggests that public health programmes should not only promote greater consumption of vegetables, fruits, legumes and aquatic products, but also emphasize individually tailored, progressive and well-supervised exercise regimens that account for age-related declines in bone and muscle strength. Targeted health education and primary care–based counselling for midlife and older women may help to optimize lifestyle behaviours, reduce the burden of LBP and support healthy ageing at the population level. Conclusions In conclusion, low adherence to a PBDP and higher physical activity levels were associated with increased odds of LBP in midlife and older women, particularly after menopause. Our findings suggest that, especially in postmenopausal women with reduced BMD or muscle strength, high mechanical loads may be associated with greater LBP risk. Rather than indicating that higher-intensity or resistance exercise should be avoided, these results highlight the need for individually tailored, progressive and well-supervised physical activity programmes that balance the musculoskeletal benefits of loading with the potential risks in vulnerable individuals. Acknowledgments Not applicable. Abbreviations Low back pain Bone mineral density Body mass index Principal components analysis Food frequency questionnaire Kaiser-Meyer-Olkin Appendix A See Tables 5, 6, 7, 8, 9. Table 5Food groupings used in the dietary pattern analysesFood groupDietary components Rice and its productsRice and its productsWheat flour and its productsWheat flour and its productsCoarse cerealsOat noodles, yellow rice noodles, buckwheat noodles, corn and its products, other cereals and products, miscellaneous beansTubersTubersBeans and their productsSoybeans, soymilk, tofu brain, tofu, bean curd type, shredded tofuVegetable mushrooms and algaeVegetables, mushrooms and algaeFruitsFresh fruit, freshly squeezed fruit and vegetable juiceMilk and dairy productsWhole liquid milk, low-fat milk, whole milk powder, milk tea, etc.PorkPorkBeef and muttonBeef and muttonPoultry, other meat products, and animal organsPoultry, other meats, meat products, animal offalAquatic productMarine fish, freshwater fish, shrimps, crabs, molluscsEggs and egg productsFresh, salted and leather eggsCooking oilVegetable oils, olive oil, linseed oil, animal oilsSalt and pickled vegetablessalt、pickled vegetableCondimentCooking sugar, soy sauce, sauces, chicken essence, monosodium glutamate, chicken powder, etc.Snacks, vegetarian desserts, nutsBread/biscuits/pastries, etc., melon seeds/peanuts/walnuts, etc., potato chips, other puffed foods, dried fruit/candied fruit, chocolate, candies, ice cream/crispy ice cream, etc., dairy-containing confections, instant noodles, etc., dried fruitFried dough foodsDoughnut, oil cake, fried cake, sesame ball, fruit, etc.Other beveragesCarbonated beverages, fruit and vegetable juice beverages, other sugar-sweetened beveragesTea and coffee drinksCoffee, tea, green tea, brick teaAlcoholic drinksHigh-altitude liquor, low-altitude liquor, beer, yellow wine, etc. Food groupings used in the dietary pattern analyses Table 6Three dietary patterns and factor loads of each food groupFood groupModel 1Model 2Model 3 PBDPNPDPNTDPFruits0.6080.140-0.038Vegetable mushrooms and algae0.6040.0280.043Beans and their products0.5030.0610.230Eggs and egg products0.4500.0370.131Coarse cereals0.431-0.0210.085Aquatic products0.3980.334-0.050Beef and mutton-0.0710.6090.142Other beverages0.1230.529-0.072Tea and coffee drinks-0.0710.5250.002Poultry, other meat products, and animal organs0.4120.5060.064Snacks, vegetarian desserts, nuts0.3510.4950.081Fried dough foods-0.0290.3690.286Alcoholic drinks0.0300.3530.013Milk and dairy products0.1800.3200Tubers0.141-0.1250.598Wheat flour and its products-0.1810.2330.575Rice and its products-0.1990.3230.544Salt and pickled vegetables0.149-0.0300.529Pork0.1880.0370.398Cooking oil0.235-0.1100.327Condiments0.1980.0980.243 Three dietary patterns and factor loads of each food group Table 7Baseline characteristics of participants by quartiles of PBDPVariables PBDPPQ1(n=314)Q2(n=369)Q3(n=357)Q4(n=490)Age0.074premenopausal women (18-44 years)17.525.726.324.5perimenopausal women (45-55 years)53.849.348.253.1postmenopausal women (≥56 years)28.724.925.522.4Nationality<0.05Han71.0 89.791.392.2Mongolian and other ethnic groups29.0 10.38.77.8Area<0.05urban10.522.527.243.7rural or pastoral89.577.572.856.3SES<0.05Low38.237.433.926.1Middle52.946.341.244.3High8.916.324.929.6BMI, Mean(SD)26.21±4.3625.40±3.5026.05±3.9825.24±3.660.001BMD (T-scores), Mean(SD)-2.30±1.83-2.28±1.732.11±1.81-1.48±1.81<0.05Sleep deficit0.104yes22.630.629.428.4no77.469.470.671.6Physical activity0.001low-intensity42.748.0 48.256.3higher-intensity57.352.0 51.843.7Sedentary time0.001<=2.0057.352.346.244.12.01-3.0019.416.518.216.93.01-5.0013.718.216.520.45.01+9.613.0 19.0 18.6Low back pain<0.05no pain81.283.288.593.7pain18.816.811.56.3 Baseline characteristics of participants by quartiles of PBDP Table 8Baseline characteristics of participants by quartiles of NPDPVariables NPDPPQ1(n=573)Q2(n=409)Q3(n=319)Q4(n=229)Age<0.05premenopausal women (18-44 years)18.820.330.133.6perimenopausal women (45-55 years)52.954.843.651.1postmenopausal women (≥56 years)28.324.926.315.3Nationality<0.05Han96.3 94.185.3353.7Mongolian and other ethnic groups3.7 5.914.746.3Area<0.05urban23.633.534.220.1rural or pastoral76.466.565.879.9SES<0.05Low42.234.524.520.1Middle42.244.345.857.6High15.521.329.822.3BMI, Mean(SD)25.68±3.5025.61±3.5125.52±4.2325.96±4.760.608BMD (T-scores), Mean(SD)-1.98±1.84-2.04±1.74-1.93±1.83-2.00±1.950.855Sleep deficit<0.05yes34.626.424.119.7no65.473.675.980.3Physical activity<0.05low-intensity54.552.3 47.635.4higher-intensity45.547.7 52.464.6Sedentary time<0.05<=2.0054.351.145.538.92.01-3.0017.316.915.423.13.01-5.0017.815.418.819.25.01+10.616.6 20.4 18.8Low back pain0.295no pain85.987.0 88.587.8pain14.113.0 11.512.2 Baseline characteristics of participants by quartiles of NPDP Table 9Baseline characteristics of participants by quartiles of NTDPVariables NTDPPQ1(n=468)Q2(n=433)Q3(n=347)Q4(n=282)Age0.982premenopausal women (18-44 years)23.724.522.824.1perimenopausal women (45-55 years)50.051.552.750.7postmenopausal women (≥56 years)26.324.024.525.2Nationality0.015Han84.6 91.287.0 84.8Mongolian and other ethnic groups15.4 8.813.0 15.2Area<0.05urban39.126.624.515.6rural or pastoral60.973.475.584.4SES0.036Low29.336.333.734.0 Middle44.743.648.447.9High26.120.117.918.1BMI, Mean(SD)25.32±3.8925.75±3.6925.59±3.8825.24±3.660.022BMD (T-scores), Mean(SD)-1.85±1.78-1.89±1.85-2.05±1.88-1.48±1.810.014Sleep deficit0.424yes30.325.926.529.1no69.774.173.570.9Physical activity<0.05low-intensity59.0 48.3 44.142.9higher-intensity41.0 51.7 55.957.1Sedentary time0.481<=2.0047.950.846.752.52.01-3.0017.117.117.319.93.01-5.0018.417.820.212.85.01+16.714.3 15.9 14.9Low back pain0.074no pain87.684.388.290.8pain12.415.711.89.2 Baseline characteristics of participants by quartiles of NTDP Appendix B See Figure 3 Scree plot of dietary pattern Author contributions X.M.W. conceived of the study, oversaw its design and coordination, provided resources, and acquired funding. H.W.L., S.W., R.X.N., P.P.W., H.L.L., and Y.L.L. contributed to the study concept and design. The methodology was developed by P.P.W., J.Z., and X.L. J.Z. and X.L. performed software development and validation, assisted by B.W.Z. J.Z. conducted data analysis, wrote the original manuscript, and created visualizations. Experimental work and data collection were carried out by Y.L.W., H.Q.Z., X.Y.W., J.Z., X.L., Q.Q.D., B.W.Z, and J.X.F. Data curation was handled by Y.L.W., H.Q.Z., X.Y.W., H.L.L., J.Z., X.L., Y.J.T., Q.Q.D., B.W.Z, and J.X.F. H.W.L. reviewed and edited the manuscript. All authors read and approved the final manuscript. Funding This study was supported by Inner Mongolia Medical University ZhiYuan Talent Program and White Paper on the Health of Ordos Residents Data availability Due to the sensitive nature of the dataset which the analysis is based upon, it is not publicly available. In special cases data is available from the authors on reasonable request. Declarations Ethics approval and consent to participate Ethical approval for this study was conducted by the Biomedical Research Ethics Committee of Inner Mongolia Medical University (No. YKD202301132). The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. All participants were in-formed about the purpose of the study, assured of confidentiality, and provided written consent prior to participation. Participation was voluntary, and respondents could withdraw at any time without consequence. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. JingZhao and HaiwenLu have contributed equally to this manuscript. References Associated Data Data Availability Statement Due to the sensitive nature of the dataset which the analysis is based upon, it is not publicly available. In special cases data is available from the authors on reasonable request.